Let Your Graph Do the Talking: Encoding Structured Data for LLMs

TL;DR

This paper presents GraphToken, a parameter-efficient method for encoding structured data into prompts for large language models, significantly improving reasoning performance across various graph tasks.

Contribution

Introducing GraphToken, the first general encoding method for structured data that enhances LLM reasoning across multiple graph-based tasks.

Findings

Up to 73% improvement on graph reasoning tasks

Explicit graph encoding enhances LLM performance

Applicable across node, edge, and graph-level tasks

Abstract

How can we best encode structured data into sequential form for use in large language models (LLMs)? In this work, we introduce a parameter-efficient method to explicitly represent structured data for LLMs. Our method, GraphToken, learns an encoding function to extend prompts with explicit structured information. Unlike other work which focuses on limited domains (e.g. knowledge graph representation), our work is the first effort focused on the general encoding of structured data to be used for various reasoning tasks. We show that explicitly representing the graph structure allows significant improvements to graph reasoning tasks. Specifically, we see across the board improvements - up to 73% points - on node, edge and, graph-level tasks from the GraphQA benchmark.